In a network environment, a modular chassis may be deployed for a group of network devices (e.g., switches or routers) to behave as a single network device. Alternatively, or additionally, a virtual chassis may be implemented by connecting a group (e.g., two or more) physically separate network devices together using cables that allow the network devices to behave as a single network device. In such networking environments, there is a high demand to upgrade the system software on these chassis without affecting the network traffic.
In a traditional dual chassis deployment, one chassis at a time can be brought down completely because there is total redundancy. Some dual chassis environments employ the use of link aggregation groups (LAGs) to provide redundancy while maximizing bandwidth by utilizing the second chassis. A LAG may be used to provide load balancing across multiple parallel links between two network devices while providing local link protection. However, LAG configurations typically distribute redundancy among all available network devices, thus preventing the bringing down of a single network device to perform software upgrades without affecting the network traffic.